“Basicity.” A Comparison of Hydrogen Bonding and ... - ACS Publications

Arnerr, et al. Comparison of H-Bonding and Proton Transfer to Lewis Bases ...... P. Haake, R. D. Cook, and G. H. Hurst, J. Amer. Chem. SOC., 89,2650 (...
0 downloads 0 Views 2MB Size
3875

“Basicity.” A Comparison of Hydrogen Bonding and Proton Transfer to Some Lewis Baseda2 Edward M. Arnett,” Edward J. Mitchell, and T. S. S. R. Murty Contribution f r o m the Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260, and the Mellon Institute, Pittsburgh, Pennsylvania 15213. Received December 26, 1973 Abstract: Hydrogen bond enthalpies (AHr’s)for the interaction of p-fluorophenol with 65 bases have been determined calorimetrically, using two independent methods wherever feasible. Heats of protonation (4H,’s)in fluorosulfuric acid for these bases have been measured also. AHr and AH, are used to compare the energetics of hydrogen bonding and proton transfer in solution, and it has been found that no single relationship exists to correlate protonation and hydrogen bonding, but that separate lines are necessary for different functional groups. If 4H, is plotted 6s. 4Hf, points representing data for basic types such as amides, phosphoroxy compounds, pyridines, sulfides, and sulfoxides fall on separate parallel lines. Solvent effects on 4Hf are discussed especially with regard to recent attempts to correct for them. AHf is correlated with various acid-base solvation parameters and we find that Gutmann’s donicity numbers, Drago’s E and C parameters, or 4 v values (the Badger-Bauer relationship) can be used to estimate reasonable AHr values, often within about 0.5 kcal/mol of our experimental results. 4Hc values and independently measured equilibrium constants for hydrogen bond formation (Kt’s) are used to consider the extrathermodynamic relations between AGro, AHr”,and 4Sf0. Neither 4 G f o nor 4St” showed any general correlation with AHi”, but some of the data could be resolved into separate trends for different groups of bases. Moreover, large changes in A G O and AHf” for pyridines, sulfoxides, amides, and phosphoroxy compounds are found to be nearly independent of entropy changes. The relation of current theories of the hydrogen bond is examined and attention is drawn to conceptual fuzziness in the definition of hydrogen-bonded systems. In conclusion, the advantages of using proton affinities in the gas phase as a primary reference point for discussing “basicity” are cited.

F

or at least 2 centuries students of chemical affinity have recognized two fundamental classes of compounds, acids and bases.3’4 Not content to define the two classes purely in terms of their mutual interaction, a long and prestigious line of natural philosophers including Lavoisier, Davy, Berzelius, Liebig, Ostwald, Arrhenius, and Br snsted has attempted to identify the essential atom or function which was inherently responsible for “acidity” or “basicity.” In 1923 G. N. Lewis proposed the unifying concept which has organized this entire field ever since. “A basic substance is one which has a lone pair of electrons which may be used t o complete the stable group of another atom. An acid substance is one which can employ a lone pair from another molecule to complete its stable grouping.”j By supplementing Lewis’ definitions with such terms as heterolysis, coordination, electrophile, nucleophile, and solcation, his concepts have been able to encompass most of the modern electronic theory of chemical reactions.6 However, despite the enormous power of the Lewis valence theory to explain and predict molecular interactions, there is still n o single quantifiable property which can be used as a general guide to “basicity” or “acidity.” In principle, if the electron density on the donor atom could be determined, this would be the ultimate measure of the inherent basicity of an isolated (1) This work has been supported by NSF Grant GP-6550-X and OSW Grant 14-30-2570for which wc are most grateful. (2) Based primarily on the doctoral thesis of Edward J. Mitchell, University of Pittsburgh, 1972. (3) F. J. Moore, “A History of Chemistry,” McGraw-Hill, New York, N.Y . , 1939, Chapter 13. (4) R . P. Bell, “The Proton in Chemistry,” 2nd ed, Cornell University Press, Ithaca, N. Y., 1973, Chapter I. (5) G. N. Leuis, “Valence and the Structure of Atoms and Molecules,” The Chemical Catalogue Co., New York, N. Y., 1923, p 142. (6) C. 1C. Ingold, “Structure and Mechanism in Organic Chemistry,” 2nd ed, Cornell University Press, Ithaca, N. Y., 1969.

Arnerr, et al.

function. Yet, if such a property were readily available, we now know (see below) that even in the gas phase, where solvation is not a complicating factor, there is no general qualitative relationship for the interaction of such Lewis acids as boron halides, the bare proton and carbenium with a variety of bases carrying different functional groups. In condensed phase chemistry, where acid-base interactions are most commonly encountered, the enormous discrepancies between the orders of reactivity of different types of acids with a given series of electron donors are even more dramatic. Various attempts have been made to classify acid-base behavior by introducing new variables. l 1 - I 3 However, recourse t o multiparameter treatments merely formalizes the fact that “acidity” and “basicity” are terms which only have operational meaning within the complex relationships of specific acid-base interactions. The transfer of a proton from one baselo (A-) to another (B) in solvent (S) is the most general and imporA-H.. .S

+ B . . . S t- B-H.. ----f

.S

+ A-.’ .S

(I)

tant reaction in chemistry.I7 Both the forward and (7) D. P. N. Satchel1 and R . S . Satchcll, Chetn. Rec., 69, 251 (1969). (8) J. L. Beauchamp, A/7nu. Rec. Pi7j.s. Chen?.,22, 552 (1971). (9) J. Franklin, Ed., ”Ion-Molecule Reactions,” Plenum Press, New York, N. Y., 1972. (10) T. B. McMahon, R. J. Blint, D. P. Ridge, and J. L. Bcnuchamp, J . Amer. Chen?. SOC.,94, 8934 (1972). (11) R. G. Pearson, J . Anzer. Chem. Soc., 85, 3533 (1963); Science, 151, 172 (1966); Cheni. Bri?., 3, 103 (1967); J. Chem. Educ., 45, 581, 643 (1968). (12) R . S . Mulliken, J . Amer. Cbem. Soc., 74, 811 (1952). (13) R. S. Drago, G. C. Vogel, and T. E. Needham, J . Anter. Chem. SOC.,93, 6014 (1971). (14) Reference 4. p 155 I?. (15) J . 0. Edwards, J . Amer. Cheni. SOC.,78, 1819 (1956). (16) The particular charge distribution on A and B i n eq 1 is of course arbitrary, but is shown i n order to emphasize that ions are made or cancelled in protolytic reactions. (17) E. M. Arnett, R . P. Quirk, and J. J. Burke, J . Amer. Chem. SOC., 92, 1260 (1970).

Comparison of H-Bonding and Proton Transfer to Lewis Bases

3876

reverse reactions must occur through transition states which are very similar to (and which may be preceded by) hydrogen-bonded complex formation. 18-z* Furthermore, the solvation of the four species shown in eq l and various subsidiary ion pairs (e.g., B-H+. . .A-) which may be present generally occurs primarily through hydrogen bonds. It is therefore natural that many authors have naively considered it obvious that a close relationship should exist between proton transfer (eq 1) and hydrogen bonding (eq 2 ) . Reaction 2 is probably a A H +B,-A-H,.,B

( 2)

good model for the initial stages leading to the transition state in eq 1. However, in eq 1 strong bonds are made and broken; and charges are created, neutralized, and solvated; so that the structural and energy differences between products and reactants must be vastly different from those represented in eq 2. The first serious attempt, of which we are aware, to correlate hydrogen bonding with proton transfer was that of Gordy and Stanfordz1 in 1940. Using Hammett‘s newly devised estimates for the pK,’s of protonated weak bases as a measure for reaction 1 they found a rather good correlation of these data with infrared frequency shifts, Av,produced by the interaction of the same bases with CH,OD. Subsequent scrutiny by weakened the case for a general correlaTamres, et UI.,~’ tion and a thoughtful study by Joris and Schleyerz3 examined the difference between Av and pK, as measures of “basicity.” In several previous publications we have drawn attention both to the problems with using Av as a criterion of hydrogen bond strength24and to the shortcomings of many pK, estimates obtained by the Hammett indicator method as reliable measures of protonation energies. l7 We have proposed the heat of protonation (AHi) in fluorosulfuric acid (HS03F) as a direct and simple quantitative measure for comparing Lewis bases in reaction 1 and, in collaboration with Professors R. Taft and P. v. R. Schleyer, have developed several methods for comparing the thermodynamics of the interaction of bases with a common hydrogen bonding acid (AH = p-fluorophenol) in eq 2. Several more recent reports?3,”6 suggested that within the limited data of our preliminary studies, major systematic differences exist between the energetics of hydrogen bonding and proton transfer for bases of different functional groups. The present article is a full report on our deterniination of heats of ionization ( A H , )in HS0,F and heats of hydrogen bonding (AHf) with p-fluorophenol (PFP) in carbon tetrachloride at 25” for 65 Lewis bases carrying a variety of functional groups. To our knowledge, this is the most extensive comparative study of hydrogen bonding and proton transfer that has yet been carried (18) J. E. Crooks and B. H. Robinson, Trans. Furadaj’ Soc., 66, 1436 (1970). (19) M. Eigen, Angew. Chem., I n t , Ed. Eiigl., 3, l(1964). (20) F. M. Jones, 111, D. Eustace, and E. Grunwald, J . Anier. Cheni. Soc., 94, 8941 (1972). (21) W. Gordy and S . C. Stanford, J. Chem. Phys., 9,204 (1941). (22) M. Tamres, S. Searles, E. Leighly, and D. Mohrman, J . Anier. Cheni. SGC.,7 6 , 3983 (1954). (23) L. Joris and P. v. R. Schleyer, Tetrahedron, 24, 5991 (1968). (24) E. M. Arnett, L. Joris, E. Mitchell, T. S . S . R. Murty, T,M. Gorrie, a n d P. v. R . Schleyer, J . Amer. Cheni. Soc., 92, 2365 (1970). ( 2 5 ) R. W. Taft, D. Gurka, L. Joris, P. v. R. Schleyer, and J. W. Rakshys, J . Anier. Cheni. Soc., 91, 4801 (1969). (26) E. M. Arnett and E. J. Mitchell, J. Anier. Chem. Soc., 93, 4052 (1971).

Journal of tile Americun Chemicul Society j 96:12

out. The relationship between the two properties will be examined and they will be compared with other “basicity” parameters in order to assess the value of current theoretical approaches.

Experimental Section Materials. All of the compounds used in this study were commercially available. Their sources, details of purification, and evidence for their purity are presented in Mitchell’s thesis.2 Generally liquids were dried, using a n appropriate agent, and then distilled through a 20-in. vacuum-jacketed column packed with glass helices and rated at six theoretical plates. If higher resolution was required, a 24-in. Nestor-Faust spinning band column was used. Solids were recrystallized t o a constant melting point (usually two or three recrystallizations) and dried in a vacuum oven or a drying pistol. Refractive indices were measured using an AbbC refractometer thermostated at 2 5 ” . Melting points were obtained using a Thomas-Hoover melting point apparatus, All compounds were purified until their properties agreed well with accepted literature values. Carbon tetrachloride (“Baker Analyzed” reagent) was distilled from calcium hydride, stored over a large bed of Linde molecular sieves (size 4A, 4 X 8 beads), and used with 2 weeks of purification. Karl Fischer titration, using a Photovolt automatic titrator, showed that random samples of the carbon tetrachloride used generally contained less than 0.001 water. Fluorosulfuric acid was purified as before. li Nuclear Magnetic Resonance Measurements. Nmr spectra were determined for fluorosulfuric acid solutions of the bases used in this study. These demonstrated that the protonation process was simple and complete; furthermore. it provided additional proof for the identity and purity of the compounds. In many cases spectra could be matched against those reported by Olah27 and Gillespie28 for the conjugate acids of the bases under stud) in HSOaF. In all cases the spectra were constant for at least a period corresponding to that required for the calorimetric measurement of A H , and could be readily interpreted in terms of the expected onium ion.2 Nmr spectra were determined with a Varian A-60 nmr spectrometer. All chemical shifts (6) were reported in parts per million (ppm) downfield from TMS at 60 MHz, using internal CH2CI2(6 5.30 ppm) as a secondary standard in fluorosulfuric acid. Samples of the protonated bases for nmr observation were prepared by slowly adding the internal standard and about 50 mg of base t o a rapidly stirred solution of 0.5 ml of HSOsF in a 2-dram vial at room temperature. These solutions were then transferred quicklq cici pipet t o an nmr sample tube whose plastic cap had been punctured to prevent buildup of pressure. The observation of simple conversion to protonated species under conditions used for the nmr studies (0.1-1.0 M ; 40”) is taken as good evidence for similar behavior under the conditions of the M ; 25‘). calorimetric measurements (cu. Enthalpy Measurements. The solution calorimeter was essentially that described originally b)’ Arnett, P I dZg Its application to hydrogen bonding (AH;) m e a ~ u r e m e n t and s ~ ~ heats of protonation (AHi)liwas as we have previously reported. Precision and Accuracy of Calorimetric Measurements. The precision of the heat measurements is dependent primarill on the magnitude cf the recorder pen deflection. The sire of this deflection itself depends upon such factors as the amount of base used, the amount of proton donor used. the AHf and Kf of hydrogen bond formation. the heat capacity of the calorimetric aqstem. and the heat of solution of the solute in the solvent. The errors involved in the measurements are reported here at the 95 %confidence level.30 When using the high dilution method (method I ) , z 4 the errors in the calculated heats of hydrogen bond formation can be as large as C.4-0.5 kcal/mol for weak bases with low Kf’s ( e . g . . 3,5-dichloropyridine; Kf = 6.4)31where the heats produced are small. I n the case of strong bases ( e . g . , phenyl methyl sulfoxide; KI = 140), (27) G . A . Olah, A . M. White, and D. H. O’Brien, Chem. Rec., 70, 561 (1970). (28) R . J. Gillcspie, Accorturs Cheni. Res., 1, 202 (1968). (29) E. M. Arnett, W. G . Bentrude, J. J. Burke, and P. McC. Duggleby, J. Amer. Chem. SOC.,87, 1541 (1965). (30) W. J. Youden, “Statistical Methods for Chemists,” Wiley, New York, N.Y., 1951. (31) R. W . Taft, private communication, June 1971.

June 12, 1974

3877 Table I. Measured Partial Molar Heats of Solution Used to Derive Enthalpies of Hydrogen Bonding in Carbon Tetrachloride (AHf) and of Ionization in Fluorosulfuric Acid (AHi). Bases in This and Table I1 Are Arranged in Order of Increasing AHi Base 1. Thionyl chloride 2. Dimethyl sulfate 3. Phosphoroxychloride 4. Diphenyl sulfide 5. Tetramethylene sulfone 6. Dichlorophenylphosphine oxide 7. Chloromethyl methyl sulfide 8. Anthrone 9. Acetonitrile 10. Phenyl methyl sulfide 11. o-Dichlorobenzene 12. Propylene carbonate 13. Diethyl chlorophosphate 14. Dimethyl sulfite 15. Diethyl carbonate 16. Cyclopentanone 17. Ethyl acetate 18. Cyclohexanone 19. Acetone 20. Triphenylphosphine oxide 21. Di-n-butyl sulfide 22. N,N-Dimethyltrifluoroacetamide 23. Diethyl ether 24. Diethyl sulfide 25. Trimethyl phosphate 26. Tetrahydrofuran 27. Tetrahydrothiophene 28. Phenyl methyl sulfoxide 29. Triethyl phosphate 30. 1,4-Dioxane 31. Diethyl ethylphosphonate 32. N-Methylformamide 33. 2,6-Dimethyl-y-pyrone 34. N,N-Dimethylchloroacetamide 35. 3,s-Dichloropyridine 36. Dimethyl sulfoxide 37. Trimethylphosphine oxide 38. 1,2-Dimethoxyethane 39. N,N-Dimeth ylbenzamide 40. N,N-Dimethylformamide 41. Di-n-butyl sulfoxide 42. Tetramethylene sulfoxide 43. Pyridine N-oxide 44. 2-Bromopyridine 45. 2-Chloropyridine 46. N-Methyl-2-pyrrolidone 47. N,N-Dimethylacetamide 48. 3-Bromopyridine 49. Quinoline 50. Pyridine 51. 1,1,3,3-Tetramethylurea 52. 4-Methylpyridine 53. 2,6-Dimethylpyridine 54. 2,4,6-Trimethylpyridine 55. Quinuclidine 56. Triethylamine 57. Phenyl phosphorodichloridate 58. Hexamethylphosphoramide

-A&”, kcal/mol Endothermic 2.69 f 0.040 4.4 f0.1 7.4 f 0 . Y 9.6 f 0 . 2 ~ 9.9 f0.3 11.4 f 0 . 2 12.7 f 0.31 12.8 f 0 . 2 12.8 f 0 . 2 14.1 f 0 . 3 14.6 f 0 . 1 14.7 f 0 . 3 14.9 f 0 . 2 16.2 f 0 . 2 17.3 f 0.11 17.4 f 0.1’ 18.1 f 0.1’ 18.3 f 0.11 18.8 f 0 . 2 18.9 f 0 . 1 19.2 f 0 . 1 19.5 f 0.78 19.5 f 0 . 3 19.6 f 0 . 1 20.2 f 0.20 20.2 f 0 . 2 20.2 f 1.0d 20.6 f 0 . 2 21.7 f 0 . 2 23.6 f 0 . 3 25.1 f 0 . 3 25.2 f 0.2, 25.5 f 0 . 2 26.4 f 0 . 3 26.5 f 0 . 2 27.0 f 0 . 6 28.3 f 0 . 2 28.4 f 0 . 2 28.6 f 0 . 1 28.9 f 0 . 5 29.1 f 0 . 1 29.1 f 0 . 3 29.2 f 0 . 3 30.7 f 0 . 3 31.3 f 0 . 4 31.6 f 0 . 1 34.4 f 0.48 35.6 f 0 . 2 38.2 f 0.28 38.2 f 0 . 7 39.1 f 0 . 3 41.0 f 0.28 42.9 f 0 . 2 45.2 f 0 . 2 49.8 f 0.18

mcc,,, kcal/mol $0.58 f 0.04 $2.59 f 0.03c +0.39 f 0.03 +0.18 f 0.03 $2.21 f 0.05 $0.84 f 0.05 1-0.45 f 0.01 + 6 . 5 f 0.4’ $1.81 f 0.10 +0.36 f 0.03 +0.37 f 0.01 $3.15 f 0.08 +0.77 f 0.02 + l . 0 8 f 0.05 +0.19 f 0.03 +0.32 f 0.02’ $0.014 f 0.004’ +0.09 f 0.011 $0.79 f 0.021 $4.15 f 0.23 -0.41 f 0.03 +1.67 f 0.04 -0.42 f 0.016 -0.54 f 0.02 $0.76 f 0.03 -0.60 f 0.026 -0.55 f 0.05 +4.96 f 0.13 +0.02 f 0.05 -0.16 f 0.01 -0.64 f 0.09 + 4 . 5 f 0.10 $6.23 f 0.24 $1.86 f 0.06 $4.28 f 0.06 $1.76 f 0.04 $5.23 f 0.09 -0.27 f 0.05 $0.71 f 0.02 +0.76 f 0.02 +0.60 f 0.09 LO.39 f 0.02 +4.28 f 0.41 $0.96 f 0.03 +0.96 f 0.04 0

+0.41 $0.23 $0.46 $0.36 -0.56 -0.05 $0.27 -0.23 $0.57 -0.64

f 0.02 f 0.036 f 0.01 f 0.028 f 0.05 f 0.01 f 0.08 f 0.03 f 0.03h f 0.08#

~~

AR, of PFP‘ in base, A P 8 of PFAb in base, kcal/mol

kcal/mol

+4.51 $3.88 +2.46 $5.15 $2.06 +1.63 +4.73

f 0.20 f 0.07 f 0.04 f 0.14 f 0.08 f 0.05 f 0.13

-0.63 $0.39 -0.14 +0.25

f 0.06 f 0.05 f 0.02 f 0.04

+4.77 +5.96 $2.01 $0.57 +2.24 $1.81 +0.14 $1.13 $0.23 $0.38

f 0.14 f 0.12 f 0.10 f 0.02 f 0.10 f 0.09 f 0.06 f 0.090 f 0.01 f 0.05

+0.13 f 0.03 +0.15 f 0.01 +0.23 f 0.04 -0.19 f 0.03

0

-0.09 f 0.02 +0.31 f 0.01

0

-0.30 -0.68 -0.46 -0.42 -0.34

f 0.02 f 0.05 f 0.040 f 0.02 f 0.03

+3.23 f 0.05

$0.36 f 0.01

+0.41 +2.77 -0.21 -0.17 +2.60

f 0.090 f 0.08 f 0.03 f 0.010 f 0.12

-0.35 +0.09 -0.08 -0.75

-0.18 f 0.03 $1.01 f 0.090 -1.95 f 0.09

$0.10 -0.22 -0.80 +0.11

00

f 0.040 f 0.02 f 0.02 f 0.010

0

f 0.03 f 0.010 f 0.06 f 0.02Q

-0.66 f 0.030

+0.22 f 0.010

-0.29 f 0.05

-0.85 f 0.02

-1.13 f 0.080

-0.49 f 0.020

-2.21 f 0.07

-0.88 f 0.04

+0.70 $0.51 -1.02 -1.76

$0.18 +0.13 -2.09 -0.65

f 0.03 f 0.03 f 0.04 f 0.110

f 0.01 f 0.03 f 0.08 f 0.01‘’

-1.14 f 0.060 -1.01 f 0.060

00 $0.06 f 0.010

-1.53 f 0.040 -2.31 f 0.07 -2.34 f 0.09

-0.27 f 0.010 -0.16 f 0.05 -0.29 f 0.05

-2.31 f 0.050 $2.12 f 0.04 -4.12 f 0.08

$0.28 f 0.010 +0.10 f 0.03 -1.71 f 0.05

PFP is p-fluorophenol. b PFA is p-fluoroanisole. c Unpublished results of Dr. John V. Carter. Compound undergoes slow exothermic reaction in HS03F. e Reference 17. f E. M. Arnett, R. P. Quirk, and J. W. Larsen, J. Amer. Chem. Soc., 92,3977 (1970). 0 Reference 24. Inert solvent used was o-dichlorobenzene. which produce larger heats, the standard deviation in the AHf values are of the order of ~tO.1-0.2 kcal/mol. The standard deviation in AH, values, using the pure base method (method I I ) , z 4 which essentially involves pooling four heats of solution3*is f0.10-0.15 kcal/mol. Wherever feasible, at least seven replica measurements of the heat of solution of the base were made in each acid solution. In several cases of particular importance using method I for AHi determination, two or three independent (32) 0. L. Davies, Ed., “Statistical Methods in Research and Production,” Hafner Publishing Co., New York, N.Y.,1957.

Arnett, et al.

repetitions were made at each acid and base concentration. Further, to verify the reproducibility of the results, some of the measurements were repeated after a time lapse of from 3 months to 1 year, using freshly purified materials. Systematic errors were avoided by frequent checks of the calorimeter against well-accepted values for the heat of solution of KC133or ethanols4in water. (33) R. J. Irvine and I. Wadso, Acta Chem. Scand., 18, 195 (1964). (34) F. Franks and B. Watson, J . Phys. E, J. Sci. Instrum., 1, 940

(1968).

Comparison of H-Bonding and Proton Transfer t o Lewis Bases

3878

Results In Table I, we present the measured thermochemical values for 58 Lewis bases necessary t o derive enthalpies of hydrogen bond formation (AHf) between p-fluorophenol and these bases in carbon tetrachloride as solvent at 25”. Corresponding data t o determine heats of protonation (AHi) in fluorosulfuric acid are also given. In Table I1 the derived results are listed and in some cases compared with independent estimates. Other relevant properties will be presented in appropriate sections of the Discussion section. The methods used t o determine AHf and AHi have been described in detail elsewhere. However, considerable experience has been gained in our laboratory and elsewhere since the methods were originally suggested and a few comments about their reliability are now in order. I. Hydrogen-Bonding Enthalpies, AHf. Most of the new data presented in this paper were obtained by one or both of the methods which we described in two previous publications. 2 4 , 35 Since one of the goals of this project was an extensive test of the consistency of the methods with each other and with reliable published values, both methods were used whenever feasible. In a few cases further comparison is made with values determined in this laboratory by Mr. John Kulluk using thermometric titration (see Discussion). Method I: The High Dilution Method. If a small quantity of a base B is injected into a dilute solution of the hydrogen bonding acid AH (or conversely, the acid is injected to a dilute solution of the base) to produce a 1 : 1 complex C, the observed heat of interaction, AHobsd, is related to the enthalpy of hydrogen bond formation (AHt) by the expression

AHobsd = AHf[C]V

(3)

where V = volume in liters of the solution in the calorimeter and [C] = molar concentration of the complex formed at equilibrium which is calculated from the equilibrium constant Kf. Our values of Kt were mainly derived from the direct determinations of our collaborators, Taft [using I9F nmr2j] and Schleyer [using infrared s p e c t r o ~ c o p y ~ ~Data ]. were treated as before.24 Provided that pure, dry materials are used, the most important experimental problem with this method is the instrumental one of detecting the effect of changing concentration on AH, B, or C in systems so dilute that the hydrogen-bonding acid AH is not associated. Very weak complexes of low Kf may require such high concentrations of AH and B in order to obtain measurable amounts of C that association of AH or B occurs or special medium effects arise. This can be a particularly difficult problem when AH{ and Kt are to be obtained from the same set of thermometric titration data.36-40 In the present study [AH] and [B] were generally varied between ( 2 and 15) X M . In no case did either (35) E. M. Arnett, T. S. S. R. Murty, P. v. R. Schleyer, and L. Joris, J . Amer. Chem. SOC.,89, 5955 (1967). (36) J. J. Christensen, J. Ruckman, D. J. Eatough, and R . M. Izatt, Thermochim. Acta, 3, 203 (1972). (37) T. F. Bolles and R. S. Drago, J . Amer. Chem. SOC.,87, 5015 (1965). (38) D. Neerinck, A. Van Audenhaege, L. Lamberts, and P. Huyskens, Nature (London), 218,461 (1968). (39) H. J. V. Tyrell and A. E. Beezer, “Thermometric Titrimetry,” Chapman and Hall, London, 1968. (40) S. Cabani and P. Gianni, J. Chem. SOC.A , 547 (1968).

Journal of the American Chemical Society

concentration exceed 0.025 M . Detailed tabulations and calculations for determination of AHf of compounds listed in Table I1 by the high dilution method are given in E. J. M.’s thesis (obtainable through University microfilms). We have shown p r e v i o ~ s l ythat ~ ~ AHt when determined by method I is very sensitive to errors in Kr especially if Kfis low. In order t o side-step the determination of Kr altogether, method I1 was previously proposed and partially tested. Method 11: The Pure Base Method. When a small increment of a hydrogen-bonding acid (AH) is injected into a large excess of pure base B, one may dichotomize the enthalpy of solution into two hypothetical termsthat due to formation of the I : 1 complex A-He . B and that due to all other thermochemically significant terms which d o not involve hydrogen bonding between AH and B. Provided that there are no strong specific association or solvation contributions to this latter term it should be approximated by the heat of solution of a model compound M which is as similar as possible in structure to AH save that it does not carry the same acidic hydrogen bonding function. Furthermore, the heats of solution for both the model compound and the hydrogen bonding acid must be referred back to a common “inert” solvent. When they are injected into the pure base as a solvent, the observed heat of solution is the sum of the heat of solution which it might be expected to give in an inert solvent and the heat due to special interaction with the base. To correct the heats of solution of the acid and the model compound, carbon tetrachloride was used as the reference inert solvent. The measurement of AH* is then a matter of determining: (1) ARs of the acid A in pure base, (2) ARB of the model M in pure base, (3) ARBof the acid A in the reference solvent, and (4)ARBof the model M in the reference solvent. Then, (AHf)AH...B = (ARBA ARBM)bess - (ARBA - A R s M ) ~ CSample ~,. calculations and further explanations can be found in references, 2 , 2 4 , 3 5 , 4 1 In the Discussion section we will comment on how well this method has stood up to scrutiny since it was originally proposed. 11. Heats of Ionization in HS0,F; AHi. The heat of protonation or ionization in fluorosulfuric acid was chosen as a widely applicable criterion of Brsnsted basicity.l? AHi corresponds simply to the heat of transfer of the base from infinite dilution in an “inert solvent” (usually carbon tetrachloride) to infinite dilution in HSOSF, i.e., AHi = A I T b ~ s o l-~ ARscct4. The method itself has been discussed extensively in the original article’? and requires no further comment here. At present, standard free energies of ionization are known for only a few very weak bases in this acidz8SO that no ASio have been determined in this medium. We will compare AHi against aqueous p K , data in Discussion section 11.

-

Discussion I. Hydrogen Bonding. A. The Determination and Use of Hydrogen-Bonding Energies. P a ~ l i n ghas ~~ (41) T. S. S. R. Murty, Ph.D. Thesis, University of Pittsburgh, 1967. (42) L. Pauling, “Nature of the Chemical Bond,” 3rd ed, Cornell University Press, Ithaca, N. Y . , 1960.

/ 96:12 / June 12, 1974

~~

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. ~. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49.

Thionyl chloride Dimethyl sulfate Phosphoroxychloride Diphenyl sulfide Tetramethylene sulfone Dichlorophenylphosphine oxide Chloromethyl methyl sulfide Anthrone Acetonitrile Phenyl methyl sulfide o-Dichlorobenzene Propylene carbonate Diethyl chlorophosphate Dimethyl sulfite Diethyl carbonate Cyclopentanone Ethyl acetate Cyclohexanone Acetone Triphenylphosphine oxide Di-n-butyl sulfide N ,N-Dimethyltrifluoroacetamide Diethyl ether Diethyl sulfide Trimethyl phosphate Tetrahydrofuran Tetrahydrothiophene Phenyl methyl sulfoxide Triethyl phosphate 1,4-Dioxane Diethyl ethylphosphonate N-Methylformamide 2,6-Dimethyl-y-pyrone

N,N-Dimethylchloroacetamide

3,5-Dichloropyridine Dimethyl sulfoxide Trimethylphosphine oxide 1,ZDimethoxyethane N, N-Dimethylbenzamide N,N-Dimethylformamide Di-n-butyl sulfoxide Tetramethylene sulfoxide Pyridine N-oxide 2-Bromop yridine 2-Chloropyridine N-Methyl-2-pyrrolidone N,N-Dimethylacetamide 3-Bromopyridine Quinoline 50. Pyridine 51. 1,1,3,3-Tetramethylurea 52. 4-Methylpyridine 53. 2,6-Dimethylpyridine 54. 2,4,6-Trimethylpyridine 55. Quinuclidine 56. Triethylamine 57. Phenyl phosphorodichloridate 58. Hexamethylphosphoratnide 59. Diphenyl sulfoxide 60. Triphenyl phosphate 61. Trimethylamine N-oxide

5 . 3 f 0.1 4.8 f 0 . 1 7.6 f 0 . 5 11.8 f 0 . 2 10.7 f 0 . 3 11.9 f 0 . 3 19.2 f 0.71 13.6 f 0 . 3 13.2 f 0 . 2 14.5 f 0 . 3 17.8 f 0 . 3 15.5 f 0 . 3 16.0 f 0 . 2 16.4 f 0 . 2 17.3 f 0.11 17.4 f 0.11 18.1 f 0.11 18.3 f 0.1f 20.9 f 0 . 3 18.5 f 0 . 1 20.9 f 0 . 1 19.5 f 0.7m 19.0 f 0 . 3 20.4 & 0 . 2 19.6 f 0 . 2 1 19.7 f 0 . 2 25.2 f 1 . 0 20.6 f 0 . 2 21.5 f 0 . 2 23.0 f 0 . 4 29.6 f 0 . 4 31.2 f 0.3f 27.4 f 0 . 3 30.7 f 0 . 3 28.6 f 0 . 2 32.2 f 0 . 7 28.0 f 0 . 3 29.1 f 0 . 4 29.5 f 0 . 2 29.5 f 0 . 6 29.5 & 0.1 33.4 f 0 . 5 30.2 f 0 . 3 31.7 f 0 . 3 31.3 f 0 . 4 32.0 f 0 . 1 34.6 f 0 . 3 1 37.0 f 0 . 2 38.6 f 0 . 3 1 37.6 f 0 . 7 39.0 f 0 . 3 40.7 f 0.3m 42.7 f 0 . 2 45.8 =k 0 . 7 49.1 f 0.2m

~

1.17 f 0.23 2.82 f 0.10 3.71 f 0.08 1.41 f 0.19 4.25 f 0.11 4.59 f 0.08 1.89 f 0.15 5 . 6 dz 0.40 4.2 f0.2

5.8 f 0.20

4.65 f 0.06h 1.67 f 0.19 0.50 f 0.14 4.53 f 0.13 5.55 f 0.07 4.07 f 0.12 4.20 f 0.10 5.50 f 0.09 4.74 f 0.120 5.66 & 0.070 5.59 i 0.08

7.4 f0.1 3.44 f 0.08 3.2 f 0 . 2 5 . 6 f 0.10 6.5 f0.1 5 . 6 f 0.10 6 . 3 f 0.1 6 . 5 f 0.1 5 . 5 f 0.19 6.9 f 0.2 6.9 f 0 . 3 5.4 f 0 . 3 6 . 6 f 0.10 7.7 f0.2

3.64

5.57 f 0.120 3.63 f 0.11 6.44 f 0.09 5.75 =k 0 . 0 8 ~ 3.71 f 0.13 6.59 f 0.09 5.10 f 0.110 7.46 f 0.12 6.44 f 0.08g

4.77 f O . l h 4.94 f 0.05;h 5.3i 7.51 4 . 19k 3.6' 5.41k 3 . 6 & 0.1" 5.71 5.30 f 0.06" 3 . 7 =t0.1m 5.96 5.0'

5.38

7.21 f 0.080 8.10

5.75 f 0.10 6.9 f 0.2 6 . 6 f 0.10 6.9 f0.1

6.97 f 0.110 7.64 f 0.11 5.83 f 0.09 5.93 f 0.07 7.38 f 0.12 7.44 f 0 . 1 3

7.09 7 . 9 f 0.5' 4.768 4.828 6.36 6.84 i O.l,h 6 . P

7.47 i 0.09. 7.40 f 0.09

7.2t 7.2t

7.59 =k 0.08s 8.36 f 0.12 8.46 f 0.11

7.3t 7.2 7.58

8.92 f 0.090 4.29 =k 0.08 8.72 f 0.110

9 . l'

7.5 f 0 . 2 7.0 f 0.2 6 . 2 f 0.20 7.35 f 0.1' 7.1 f 0 . 1 ~ 7.8 f 0 . 2 7 . 3 f 0.10 7.8 f 0 . 3 7.9 f 0.1 9.5 f0 . P 8.0 f 0.10 6 . 2 =k 0 . 3 6 . 7 =k 0 . 2 8.8 f 0 . 3 ~

5.26 6.39

8.1e

Heat of transfer of the base from CC14 to HSOaF, ABsE~Oa~ - A&cl,. b Heat of hydrogen bond formation to p-fluorophenol in CC14 measured using method I, the dilute solution method. c Heat of hydrogen bond formation to p-fluorophenol in the pure base as solvent (method 11). Literature values for the interaction of the phenol with these bases. The values can be compared directly with those for the interaction of PFP with bases (ref 41). E D. P. Eyman and R. S. Drago, J . Amer. Chem. SOC.,88,1617 (1966). f E. M. Arnett, R. P. Quirk, and J. W. Larsen, J. Amer. Chem. SOC.,92, 3977 (1970). 0 Reference 24. T. D. Epley and R. S. Drago, J. Amer. Chem. SOC.,89, 5770 (1967); R. S . Drago and T. D. Epley, ibid., 91,2883 (1969). i W. C. Duer and G. L. Bertrand, ibid., 92, 2587 (1970). i G. Aksnes and T. Gramstad, Acta Chem. Scand., 14, 1485 (1960). R. West, D. L. Powell, M. K. T. Lee, and L. S . Whatley, J. Amer. Chem. Soc., 86, 3227 (1964). ' R. L. Middaugh, R. S. Drago, and R. Niedyielski, J. Amer. Chem. SOC.,86,388 (1964). Reference 17. n G. C. Vogel and R. S. Drago, J. Amer. Chem. SOC.,92, 5347 (1970). These authors maintain that this value is "too low" due to an exothermic interaction between the solvent (CCI,) and the base. It should be noted, however, that their value is in close agreement with our value which was obtained using a method in which the base and CCL do not come in contact. 0 M. D. Joesten and R. S . Drago, J. Amer. Chem. SOC.,84, 3817 (1962). * D. Neerinck, A. van Audenhaege, and L. Lamberts, Ann. Chim. (Paris),4,43 (1969). g R. S. Drago, B. Wayland, and R. L. Carlson, J. Amer. Chem. SOC.,85,3125 (1963). T. Kubota, J. Amer. Chem. SOC.,88,211 (1966). 8 T. Kitao and C. H. Jarboe, J. Org. Chem., 32,407 (1967). T.Gramstad, Acra Chem. Scand., 16,807 (1962). u Solvent used was o-dichlorobenzene. y S . Singh and C. N. R. Rao, Can. J . Chem., 44, 2611 (1966). Solvent used was &chloromethane. 2 B. Styme, H. Styme, and G . Wettermark, J. Amer. Chem. Soc., 95,3490(1973).

Arnett, et al.

1 Comparison of H-Bonding and Proton Transfer to Lewis Bases

3880

provided a useful description of the hydrogen bond by stating “that under certain conditions an atom of hydrogen is attracted by rather strong forces to two atoms-so that it may be considered to be acting as a bond between them.” During the past 4 or 5 decades an enormous literature has d e ~ e l o p e d ~ ~documenting -~’ intensive efforts t o define exactly the “conditions” for this attraction, t o quantify these “rather strong forces” and to understand the nature of this peculiar bond in quantum mechanical terms. At its lower limit the strength of the hydrogen bond shades into the weak forces (dispersion, dipole-dipole, etc.) which operate between all molecules which carry covalent bonds t o hydrogen. At its upper limit, attraction of a covalent hydrogen for a second atom becomes so great that proton transfer occurs. Ordinary hydrogen bonds between neutral oxygen and nitrogen compounds have energies between 3 and 10 kcal/mol. Because of their great importance t o many areas of applied and biological chemistry, particular effort has gone into attempts to isolate and measure the strengths of hydrogen bonds between the common organic functions bearing acidic hydrogen and those with basic lone pair electrons. It is important t o appreciate that the use to which hydrogen bonding data may be put by applied chemists or biochemists is quite different from that by theoreticians. The former will have a greater need for knowing the interaction energies of somewhat complex model compounds in condensed phases-the latter usually prefer simplified systems in the gas phase. For the most part our approach has been the former one although as we shall see below, solvent effects on hydrogen-bonding energies are mostly small and probably can be estimated within 5 %in most cases. It is also important t o realize that there is both an experimental and even conceptual lower limit below which it becomes a matter of taste whether or not one chooses to say a “hydrogen bond” exists between two molecules. Thus if we consider the complex between chloroform and trimethylamine and progressively replace the chlorines on the former atom with hydrogens until it is methane, we pass by stages from a system with a fairly strong hydrogen bond to one without one. At what stage does the hydrogen bond become a normal weak dipole-dipole attraction; in CHKI, or CH3Cl? As a n operational thermochemical definition we have arbitrarily chosen t o say that a hydrogen bond exists between A-H and B if under similar conditions the enthalpy of interaction is at least 1 kcal/mol greater than the comparable interaction which is found between B and A-X where A-X is the best possible model for A-H if the A-H bond were a normal C-H bond incapable of engaging in specific intermolecular hydrogen bonding.48 The use of the “pure base method” clearly (43) (a) G. C. Pimentel and A. L. McClellan, “The Hydrogen Bond,” W. H. Freeman, San Francisco, Calif., 1960; (b) Annu. Reo. Phys. Chem., 22,347 (1971). (44) D. Hadzi, Ed., “Hydrogen Bonding,” Pergamon Press, New York, N. Y., 1959. (45) S . N. Vinogradov and R. H. Linnell, “Hydrogen Bonding,” Van Nostrand-Reinhold, New York, N. Y., 1971. (46) P. A. Kollman and L. C. Allen, Chem. Reo., 72,283 (1972). (47) A. S. N. Murthy and C. N. R. Rao, Appl. Specrrosc. Reo., 2, 69 (1968). (48) The problem of defining hydrogen bonding energy is thus equivalent to that for aromaticity and depends on the model one chooses to represent behavior of the system in the absence of H-bonding (or aromaticity).

Journal of the American Chemical Society

involves the appropriate choice of such models and also assumes that solvent effects on hydrogen bonded systems should be small. We will now consider the present status of both of these matters. 1. Solvent Effects on AHr. Even if one is ready to say that all of the interaction energy between AH and B in the complex A H . B is due to hydrogen bonding considerable care must still be exercised in determining the energy of forming this complex in solution. The generally recognized conditions for such a careful study are that neither AH nor B should be associated to an energetically significant degree (no solute-solute interactions) and that there should not be significant solventsolute interactions between the medium and AH or B which are peculiar to that solvent. (Chloroform would be a poor solvent for studying the interaction of phenols with bases.) Therefore the solvent must interact strongly enough with AH and B to dissociate their molecules, but must not d o this through strong specific interactions. These and related problems are described clearly by Drago, et aZ.49 If specific solvation forces have to be overcome in the formation of A H . . . B from AH and B, this would introduce a spurious endothermic term causing the stability of the complex t o be underestimated. Ultimately there will always be some interactions which are idiosyncratic to any solute-solvent pair; our problem is not so much that of identifying specific interactions but of deciding on how large they must be to deserve special attention. Despite the importance of the hydrogen bond there have been few extensive systematic studies of solvent effects on hydrogen bond energies. Allerhand and Schleyerso have reported a careful investigation of solvent effects on several infrared 0-H . * 0 frequencies in 21 nonbasic solvents and the gas phase. They found that medium effects were sometimes quite large and did not follow current theories of solvent shifts.50 Osawa and Yoshida obtained similar results with fewer solvents and concluded that the solvent shifts were mainly ascribable to dipole-dipole forces. A reasonable approach for relating hydrogen bonding energies in different solvents has been advanced by Christian and his colleague^.^^-^^ They propose that the energy of transfer of A-H . . B from one solvent t o another is proportional to the sum of the transfer of AH and B. This has not been tested adequately but comparison of his association constants for the pyridine-water complex in various with Allerhand’s and Schleyer’s G valuessashows the same qualitative ordering. Four values of log K for pyridine-Is complex (Figure 1) even give a close linear correlation with G. For the present purposes a useful application of Christian’s equation is seen in Figure 2 where heats of hydrogen bonding of phenol (see Table 111) with various bases in cyclohexane are plotted against AHr for the same complex in carbon tetrachloride as solvent.

-

+

+

+

(49) R. S. Drago, M. S. Nozari, and G. C. Vogel, J. Amer. Chem. Soc., 94, 90 (1972). (50) A. Allerhand and P. v. R. Schleyer, J . Amer. Chem. Soc., 85, 371 (1963). (51) F. Osawa and 2.Yoshida, Spectrochim. Acta, 23,2029 (1967). ( 5 2 ) S. D. Christian, J. R. Johnson, H. E. Affsprung, and P. J. Kilpatrick, J . Phys. Chem., 70, 3376 (1966). (53) S. D. Christian, K. 0. Yeo, and E. E. Tucker, J. Phys. Chem., 75, 2413 (1971). (54) S. D. Christian, R. Frech, and I(. 0. Yeo, J. Phys. Chem., 77, 813 (1973). (55) S . D. Christian, J. Amer. Chem. SOC.,91, 6514(1969).

/ 96:12 1 June 12, 1974

3881 9.0I

Y

1

Hexane 0 Cyclohexane

2s2/ 2.1

-

CI)

A

2.0

Carbon Tetrachloride 0 Carbon Disulfide

I

1.9-

I

I

I

I

I

:

"O

4.0

5.0

6.0

-*

7.0

8.0

9.0 10.0

(solvent)

Hf(teollmolr,

Figure 3. Heats of hydrogen bond formation in various solvents us. heats measured in benzene; numbers refer to Table IV; (0) AHr in 1,Zdichloroethane; (8)AHr in o-dichlorobenzene; ( 0 ) AHrin carbon tetrachloride; (A) AH1 in cyclohexane.

8.0-

-I e

7.0-

5 - 6.0" -

st

7

' 9 .OO

5.0-

4.01 3.0

/

2'o!,0

310

'

O

i

i

t\ 410

5!0

610

710

20

-AHCC'k (kc4lmole)

Figure 2. Heats of hydrogen bond formation measured in cyclohexane us. heats measured in CCla;numbers refer to Table 111. Table III. Solvent Effects on Heats of Hydrogen Bonding between Phenol and Various Bases

\ '0

1. 2. 3. 4.

Pyridine THF Acetone 1,4-Dioxane 5. Tetrahydrothiophene 6. Diethyl sulfide

8 . 0 0 ~ t0.1 6.7 f 0.1 6 . 6 & 0.1 6.0 i 0.05 4 . 9 i:0 . 1 4 . 6 =t0 . 1

6 . 6 f 0.2 5.0 f 0.2 4.8 i 0 . 2 4.4 f 0.1 3 . 7 i: 0 . 1 3.6 f 0.1

a a

b b

Figure 4. Heats of hydrogen bond formation in various solvents us. G values; numbers refer to Table IV: (0)AH{ in 1,Zdichloroethane; ( 0 ) AHf in carbon tetrachloride; (A) AH1 in cyclohexane; (m) AHf in benzene.

b b

L. Lamberts, 2.Phys. Cliem. (Frankfurt am Main), 73, 159 (1970). G. C . Vogel and R. S . Drago, J. Amer. Chem. SOC.,92, 5347 (1970).

Figure 3 uses data from Drago'sS6 work (see Table IV) to illustrate the same point. In Figure 4 enthalpy data from the work of Nozari and Drago are plotted against Schleyer's G values. Some of the strongest evidence that solvent effects on hydrogen bonding energies are usually modest and systematic is provided by the rather good success of the pure base method described below. TaftZ5has also noted the small effect of solvent effects on Kffor a few cases of hydrogen bonds from hydroxyl acids. 2. Present Status of the Pure Base Method (See Re(56) M. S . Nozari and R. S . Drago, J. Amer. Chem. Soc., 94, 6877 (1972). An extension of this work appeared after the present article was submitted: R. M. Guidry and R. S. Drago, J. Phys. Chem., 78, 454 (1974). For a very recent test of criticisms by these authors, see G. Olofsson and I. Olofsson, J . Amer. Chem. Soc., 95, 7231 (1973).

Amett, er al.

sults for a Restatement of the Method). This simplistic approach for estimating hydrogen bonding enthalpies was proposed by us24,35,41 to deal with cases where Kf was so small or hard to obtain that the high dilution method would lead to gross errors in AHt. We were skeptical of its reliability. 2 4 After preliminary tests with about ten bases whose AHt values had been reliably determined by other methods, we were pleasantly surprised t o find that it worked rather well. Compounds with large Kf'swill usually be highly polar or polarizable and thus should impose relatively large solvent effects on AHt so that one would expect the pure base method t o be at its worst in just those circumstances where large Kf'smake the high dilution method to be at its best and there is no need for the pure base method. Thus the two methods complement each other and could hardly be expected t o match exactly across a wide range of Kt's or AHt's. Nonetheless, for 15 compounds in Table I1 the average difference between the methods is 0.4 kcal/mol in our hands and the average difference

/ Comparison of H-Bonding and Proton Transfer to Lewis Bases

3882 Table IV. Solvent Effects on Heats of Hydrogen Bonding between m-Fluorouhenol and Various Bases“

-AHfcyclahexane

Base 1. 2. 3. 4. 5.

Ethyl acetate Dimethyl sulfoxide Pyridine Triethylamine n-Butyl ether

6 . 7 f 0.1 8 . 4 f 0.1 9.8 f 0 . 2 6.5 f 0.1

-AHfCC14

-AHfo-dichloroben~ene

5 . 2 zt 0 . 1 7 . 2 f 0.1 7 . 5 =!= 0.1

4.7 f 0.1 6 . 7 zt 0.1 6 . 9 d= 0.1 9.3 f 0.1 5.7 f 0.2

6 . 0 f 0.1

-AHfbenzene

4 . 0 f 0.1 6.1 f 0 . 1 6 . 3 f 0.1 8.6 f 0 . 1

_AHfl~2-dichloroethsae

3.7 f 0.2 5 . 4 f 0.1 6.4 j z 0 . 1 8 . 8 f 0.1 4.5 f 0.2

All data taken from ref 56. 0.0

2

I

-

$

I

-

A

U

U

0.2 0.4 0.6 0.8 1.0 Mole Fraction of the Base in Carbon Tetrochloride

0

Figure 5. A plot of mole fraction of base in carbon tetrachloride us. - A H f in kcal/mol a t each concentration: (1) diethyl ether as the base, -AH*” dilution = 5.6 f 0.1 kcal/mol; (2) N-methylformamide, 5.5 i. 0.1 kcal/mol; (3) dimethyl sulfoxide, 6.7 i 0.1 kcal/mol; (4) quinoline, 7.35 i 0.1 kcal/mol; (5) tetrahydrofuran, 5.6 f 0.1 kcal/mol; (6) dimethylformamide, 6.6 i 0.1 kcal/mol; (7) pyridine, 7.1 f 0.1 kcal/mol; (8) 4-methylpyridine, 7.3 f 0.1 kcalimol.

emphasized in the inert solvents. Lambertsp3 showed in the cases of acetone, dioxane, and pyridine (see below) that AH, values relative to tetrahydrofuran were not only insensitive to the choice of model compound (anisole US. toluene) but also to the choice of solvent (carbon tetrachloride us. cyclohexane). For many applications in solution chemistry especially in comparing the strengths of hydrogen bonds in biological systems absolute estimates of A H f are not needed.64 In such cases the evidence so far available indicates that the heat of transfer for the acid from one pure base to another corrected by the heat of transfer of a reasonable model compound (of size and shape similar to AH) should give an estimate of relative AH* within 5 %. A means for improving the agreement between the pure base method and high dilution method is shown in Figure 5.‘jj The combined heats of transfer for PFP and PFA from pure base are plotted for a series of binary mixtures of the base with carbon tetrachloride and the resulting line extrapolated to infinite dilution of the base. This approach can only be used for systems which are completely complexed across the range of solvent concentrations. If K f is low and complexing is weak, a curve rather than a straight line will result. By using the pure base AHf value and eq 3 it is possible to estimate [C] at various points on the curve and hence Kf, Iteration gives values for Kr and AH, which agree rather well with those derived by other means.41 Table V shows that for the bases listed, the extrapolated inter-

for 29 compounds between our pure base values and literature values (for phenol) is 0.5 kcal/mol. Most of Table V. Comparison of AHf’s for PFP with Various Bases the comparisons are for hydrogen bonds above 5 kcal/ Using Dilute Solution Method and the Pure Base Method Extrapolated to Infinite Dilution mol in strength. The most serious discrepancies are for the formamides which are very polar or for several - AHf, kcal/mol pyridines whose Kf’s are low and dubious, Base a b C Many polar Lewis bases have a tendency to associate 7.40 f 0.09 Pyridine 7 . 2 +C 0 . 2 7.1 =t 0.1 through dipole-dipole interactions in “inert” solvents 7 . 5 =t 0 . 2 7.35 +C 0 . 1 7 . 4 7 & 0.09 Quinoline such as carbon tetrachloride or h y d r o ~ a r b o n s . ~ ~ - ~4-Methylpyridine l 7.59 f 0 . 0 8 7.4 f 0 . 2 7.3 f0.1 5.75 & 0.08 Tetrahydrofuran 6 . 0 =t 0 . 3 5 . 6 i 0 . 1 This can introduce errors from solute-solute interactions 5.57 i 0.12 5 . 6 f 0.3d 5 . 6 i 0 . 1 Diethyl ether in applying high dilution methods to bases of low Kf. 7.21 =k 0.08 Dimethylsulfoxide 6 . 8 f 0 . 2 6 . 7 i 0.1 Bertrand and Duers2 have suggested that errors in the 6.97 i 0.11 Dimethylformamide 6 . 7 f 0 . 2 6 . 6 f 0.1 pure base method caused by association or a poor choice a Dilute solution method (method I ) . 2 4 * Pure base method of the model compound can be considerably reduced by extrapolated to infinite dilution. Values in pure base (method eliminating the use of an inert reference solvent and II).*4 d Kf used for this base is determined from calorimetric data comparing directly the heats of transfer of A-H and alone. Kf’s used for the other bases are measured by ir.24 the model compound from one pure base to another. Apparently differences between model compounds are cept is within experimental error of AHf as determined (57) W. Partenheimer, T. D. Epley, and R. S . Drago, J. Amer. Chem. CC14 by the high dilution methods. Table VI conin Soc., 90, 3886 (1968). (58) M. Rabinowitz and A . Pines, J . Chem. SOC.B, 1110 (1969). firms that the high estimates of A H f obtained for these (59) T. F. Lin, S. D. Christian, and H. E. Affsprung, J . Phys. Chem., bases by the pure base method are related to the polarity 71, 968 (1967). of the medium. (60) J. H. Hildebrand and R. L. Scott, “The Solubility of Nonelec7

trolytes,” Dover Publications, New York, N. Y., 1964, Chapter XI. (61) J. S. Rowlinson, “Liquids and Liquid Mixtures,” Butterworths. London, 1959. (62) W. C. Duer and G. L. Bertrand, J. Amer. Chem. Soc., 92, 2587 (1970).

Journal of the American Chemical Society J 96:12 J June 12, 1974

(63) L. Lamberts, Z . Phys. Chem. (Frankfurt am Main),7 3 , 159 (1970). (64) I. M. Klotz and S. B. Farnham, Biochemistry, 7,3879 (1968). (65) This section is taken from ref 41.

3883 Table VI. Estimate of Solvent Effects on AH{ for PFP as Proton Donor with Various Bases

Base

Dielectric constant at 25"

-AHf in pure base, kcal/mol

-AH{" at infinite dilution (extrapolated), kcal/mol

- AAHf, kcal/mol

Diethyl ether Tetrahydrofuran Quinoline Pyridine Dimethylformamide Dimethyl sulfoxide N-Methylformamide

4.22 7.39 9.22 12.3 36.7 48.9 182.4

5.57 f 0.12 5.75 f 0.08 7.47 f 0.09 7.40 f 0.09 6.97 f 0.10 7.21 f 0.08 6.44 f 0.08

5.6 f 0.1 5.6 f 0.1 7.35 f 0 . 1 7.1 f 0 . 1 6.6 f 0.1 6 . 7 f 0.1 5.5 f 0 . 1

0 0.15 f 0.13 0.12 f 0.14 0.30 f 0.14 0.37 f 0.14 0.41 f 0.13 0.94 f 0.13

Table VII. Thermodynamic Parameters for the Hydrogen-Bonded Complex of PFP with Various Bases Base 8. 9. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 28. 29. 32. 33. 34. 35. 36. 37. 39. 40. 41, 43. 44. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55.

56. 58. 59. 60. 61. 62. 63. 65. 66.

Anthrone Acetonitrile Ethyl acetate Cyclohexanone Acetone Triphenylphosphine oxide Di-n-butyl sulfide

N,N-Dimethyltrifluoroacetamide Diethyl ether Diethyl sulfide Trimethyl phosphate Tetrahydrofuran Phenyl methyl sulfoxide Triethyl phosphate N-Methylformamide 2,6-Dimethyl--ppyrone

N,N-Dimethylchloroacetamide 3,5-Dichloropyridine Dimethyl sulfoxide Trimethylphosphine oxide N,N-Dimethylbenzamide N,N-Dimethylformamide Di-n-butyl sulfoxide Pyridine N-oxide 2-Bromopyridine N-Methyl-Zpyrrolidone N,N-Dimethylacetamide 3-Bromopyridine Quinoline Pyridine 1,1,3,3-Tetramethylurea 4-Methylpyridine 2,6-Dimethylpyridine 2,4,6-Trimethylpyridine Quinuclidineh Triethylamine Hexamethylphosphoramide Diphenyl sulfoxide Triphenyl phosphate Trimethylamine N-oxidei 4-Dimethylaminopyridine 2-Butanone Cyclopropylamine NjV-Dimethylamino-3propionitrile

Kf 2@, l./mol 17.6 f 0.2* 7.50 12.3 f 0.4b 20.5 f 0.9 15.46 1456 f 806 1.86 26. 10.3 f 1 . P 1.30 2960 17.7 f 0.5' 1406 5406 90 f 6b 318' 48" 6.90 346 f 8b 3090e 166. 116 f 3b

Kf sw,o 1.1 mol

442 128

400b 5768 8.7= 235e 260 f 12b 20.3 f 0.96 72.3 f 1.0b 76.2 f 1.1' 1860 109 f 5b 135" 200' 4200 85.2 f 1 . 9 36W 105 f 1' 54' 3680i 407ftk 15.6 f 0 . 9 44 f 2 b

111

17.1 f 0.9'

-AGf

O,

kcal/mol

1.70 f 0.01 1.43 1.49 f 0.02 1.79 & 0.02 1.60 4.32 f 0.01 0.36 1.60 1.38 f 0.05 0.15 3.37 1.70 f 0 . 2 2.92 3.73 2.67 f 0 . 3 3.40 2.68 1.14 3 . 4 6 f 0.02 4.74 3.02 2.81 f 0.01 3.53 3.75 1.28 3.22 3.29 f 0.03 1.78 f 0.03 2.54 f 0.01 2.56 f 0.01 3.10 2.78 f 0.03 2.91 3.13 3.58 2.63 f 0 . 2 4.85 f 0.02 2.76 f 0.01 2.35 4.84 3.56 1.63 f 0.02 2.24 f 0 . 1 1.68 f 0 . 3

-AHf

O,

kcal/mol

-AH{"," kcal/mol

5 . 6 f 0.4'

4.2 f 0.20 4.74 f 0.12b>d 5.8 f 0.2b 5.59 f 0.08d 7 . 4 f O.lb 3.44 f 0.08d 3.2 f 0.2' 5 . 6 f 0.lb 3.63 f O.lld 6.5 f 0.1' 5.6 f 0.1' 6 . 3 f 0.11 6.5 f 0.1, 5 . 5 f 0.1* 6.9 f 0.2' 6 . 9 f 0.3' 5.4 f 0.4f 6 . 6 f O.lb 7.7 f 0.2' 6 . 9 f 0.2' 6 . 6 f 0.1' 6 . 9 f 0.1' 7 . 5 f 0.2' 5.83 f 0.09d 7 . 0 f 0.2' 7.44 f O.llbd 6 . 2 f 0.2* 7.35 f 0.1c 7 . 1 f 0.1' 7 . 8 f 0.2, 7.3 fO.l* 7.8 f 0.3' 7.9 f 0.1' 9 . 5 f 0.2' 8.92 f 0.09*sd 8 . 0 f O.lb 6.2 f 0.3b 6.7 f 0.2' 8.8 f 0.3' 8 . 4 f 0.2' 5.20 f 0.13bSd 7 . 5 f 0.30

4.7 5.5

5.3

6.0 6.10

7.10 6.6 6.9

8.3

5.8 f 0.1'

-A&

O,

eu

13.1 i 0 . 1 9.3 f0.7 10.9 f 0 . 3 13.4 f 0.7 13.4 f 0 . 3 10.3 i 0 . 3 10.3 f 0 . 3 5.4 i0.7 14.2 f 0 . 3 11.7 f 0 . 3 10.5 f 0 . 3 13.1 f 0 . 3 11.2 f 0 . 3 9.4 i0.3 9.5 f 0.3 11.7 i 0 . 7 15.5 f 1 . 0 14.3 f 1 . 3 10.9 i 0 . 3 10.1 f 0 . 7 13.0 & 0 . 7 12.7 i 0 . 3 11.3 it 0 . 3 12.6 i 0 . 7 15.3 f 0 . 3 12.7 i 0 . 7 13.9 f 0 . 3 14.8 f 0 . 7 16.1 f 0 . 3 15.2 f 0 . 3 15.8 i 0 . 7 15.2 i 0 . 3 16.4 i 1 . 0 16.0 f 0 . 3 19.8 f 0 . 7 21.0 f 0 . 3 10.6 f 0 . 3 11.5 i 1 . 0 14.6 f 0 . 7 13.3 i 1 . 0 16.2 f 0 . 7 12.0 =k 0 . 3 17.6 f 1 .O 13.8 f 0.3

~~~

AHf determined by 4 Determined by Mr. John Kulluk using a thermometric titration technique.86 b Reference 24. c Referenee 41. Reference 25. AHf determined by the dilute solution method (I). 0 Reference 31. o-Dichlorobenzene the pure base method (IIj. Dichloromethane was used as solvent. i T. Kubota, J. Amer. Clrem. SOC.,89,459 (1967). This value was revised was used as solvent. by Taft. from that reported in ref 25, and AGf", AH,', and A&" have been recalculated to conform with the revised equilibrium constant. They supplant the values reported in Table V of ref 24.

'

3. Concerning the Choice of Solvent for HydrogenBonding Studies. Were it not for the fact that most important hydrogen bonding energies are in the range of 5-8 kcal/mol, solvent effects of &0.5 kcal/mol could be considered insignificant. We have seen above that for condensed phase comparisons the pure base method Arnett, et al.

gives good estimates of relative AHf's and Figures 1-4 suggest that through suitable solvent parameters it may be possible to make some corrections for medium changes, although the data correlated here are too few to be more than suggestive. Drago and his colleaguesss have employed substitu-

1 Comparison of H-Bonding and Proton

Transfer to Lewis Bases

3884

A 56

37

A20

*17 99

022

BertrandB2do not show it to be unique. Until there is direct objective evidence for large abnormal interactions between pyridine and carbon tetrachloride we see no reason to set it apart from other common systems since many of them also show singularities of 0.3 kcal/mol or more. B. Thermodynamic Parameters for HydrogenBonded Complexes of PFP ; Extra Thermodynamic Relations. Relations between AGr O, A&', and AHf Table VI1 lists the thermodynamic parameters calculated for the hydrogen-bonded complexes of p-fluorophenol with 44 bases. This extends and updates our previous tabulation. 2 4 The values for the standard free energy of hydrogen bond formation to PFP, AGfO, have been calculated from the spectrophotometric equilibrium constants determined by our collaborators.24~25 Consequently, the AH,' and ASf" data listed in Table VI1 are based on essentially independent enthalpy and free energy measurements. Therefore, their correlationG8can be attempted without much likelihood6g that we might only be generating the error contours of two sets of numbers (AH', AGO, or ASo) obtained from the same set of experimental data. None of the data presently available from other published studies of which we are aware is based on entirely independent determinations of AGEo and A H f ofor the estimation of ASfo. The great majority has employed the van't Hoff equation to determine AHf and ASfO from the temperature coefficient of Kr (or AGO). In the last decade, the development of solution calorimetry, particularly titration calorimetry, has generated a variety of approaches for estimating Kf and AHf from the concentration dependence of the observed heat of interaction between titrants-in this case AH and B.36-338,41Both the van't Hoff method and the thermometric titration methods are somewhat vulnerable in principle since the same data are being used to determine two variables so that errors can develop in both. 39,40 We have previously noted good agreement between our results using independent AGtO and AHr' with high quality van't Hoff data and with Drago's thermometric titration data. In Table VI1 several measurements determined in our laboratory by Mr. John Kulluk using the continuous thermometric titration technique of Izatt and C h r i ~ t e n s e nare ~ ~ presented for comparison and give support for the validity of both methods (provided, of course, that concentrations are low and other experimental requirements are met). The results presented here for 44 bases with a single hydrogen bonding acid are the most extensive set of such data yet available and generally support OUT previous conclusions regarding the relationships of AGr O, AHt, and ASf O on half as many compounds. Thus, large changes in AGf' and AHfofor pyridines, sulfoxides, amides, and phosphoroxy compounds are nearly independent of entropy changes. The value of AS,' is essentially constant within each of these classes of bases. This leads to the behavior shown in Figure 6, where AGf' and AHf' are correlated by a series of crudely parallel lines, each of about unit slope, correlating free energy with enthalpy for the pyridines, sulfoxides, phosphoroxy compounds, and amides. For

2 10

1

2

3

4

5

-AGF in kcol/mole

Figure 6. Plot of A H f " for PFP with various bases us. AGf' for PFP with the bases; numbers for data points correspond to numbers listed in Table VII, e.g., data point 8 refers to anthrone.

tion reactions in order to eliminate the contribution from AH in various solvents in a manner similar to that used by Bertrand and Duer in pure bases.62 Using the complexing of m-fluorophenol with five Lewis bases (DMSO, ethyl acetate, di-n-butyl ether, pyridine, and triethylamine) in five nonpolar solvents they have found that the majority of the 30 displacements studied by them met their criteria of being "solvation free" within f0.2 kcal/mol. In some cases, such as the aggregation of amides in cyclohexane, there is a reasonable model for discrepancies. In others, such as those where 1,2dichloroethane is the solvent, the results did not fit their model and unexplainable factors were inferred. In our opinion it is not surprising that such discrepancies arise or that they are unexplainable. Even if we assume that all of the reported enthalpies of complexing and solution are both accurate and precise to within 0.1 kcal/mol, we do not see how it is possible to account for or predict solvent-solute interactions for these diverse systems within 0.3 kcal/mol of the theoretical gas phase value. We even doubt that it is physically significant to assign the gas phase interaction energy between the common organic bonding acids and bases completely to the "hydrogen bond" within 0.3 kcall mol considering the conceptual looseness of the term to which we have referred. Finally we would like to respond briefly to criticisms have raised regarding the which Drago, et uZ.,66n67,i3 use of carbon tetrachloride to study the hydrogen bonding to pyridines. Referring originally to the study of Morcom and T r a v e r P who determined the enthalpy of complexing of pyridine with this solvent as 0.3 kcal/ mol, Drago and his colleagues have gradually raised this to 0.9 kcal/m01~~ in order to account for failures of this system to fit their correlations. By similar indirect methods Purcel16i has suggested that the interaction energy is 1.9 kcal/mol. As can be seen from Figures 1-5 there is nothing especially abnormal about pyridine in carbon tetrachIoride compared to other solvent-base systems. Furthermore, the results of LambertsG3and (66) K. W. Morcom and D. N. Travers, Trails. Faraday Soc., 62. 2063 (1966). (67) A. D. Sherry and I